Added context_find_prefix().

[nmh] / uip / mhfixmsg.c

diff --git a/uip/mhfixmsg.c b/uip/mhfixmsg.c
index 8d67e827b9e933cfa9034187652a2dbb76975eda..d3c23767b8928c7a4ce822d2236a41b0423c7815 100644 (file)
--- a/uip/mhfixmsg.c
+++ b/uip/mhfixmsg.c
@@ -1,5 +1,5 @@
 /*
 /*

- * mhfixmsg.c -- rewrite a message with various tranformations
+ * mhfixmsg.c -- rewrite a message with various transformations

  *
  * This code is Copyright (c) 2002 and 2013, by the authors of nmh.
  * See the COPYRIGHT file in the root directory of the nmh
  *
  * This code is Copyright (c) 2002 and 2013, by the authors of nmh.
  * See the COPYRIGHT file in the root directory of the nmh


@@ -16,6 +16,9 @@
 #define MHFIXMSG_SWITCHES \
     X("decodetext 8bit|7bit", 0, DECODETEXTSW) \
     X("nodecodetext", 0, NDECODETEXTSW) \
 #define MHFIXMSG_SWITCHES \
     X("decodetext 8bit|7bit", 0, DECODETEXTSW) \
     X("nodecodetext", 0, NDECODETEXTSW) \

+    X("decodetypes", 0, DECODETYPESW) \
+    X("crlflinebreaks", 0, CRLFLINEBREAKSSW) \
+    X("nocrlflinebreaks", 0, NCRLFLINEBREAKSSW) \

     X("textcharset", 0, TEXTCHARSETSW) \
     X("notextcharset", 0, NTEXTCHARSETSW) \
     X("reformat", 0, REFORMATSW) \
     X("textcharset", 0, TEXTCHARSETSW) \
     X("notextcharset", 0, NTEXTCHARSETSW) \
     X("reformat", 0, REFORMATSW) \


@@ -80,11 +83,15 @@ typedef struct fix_transformations {
     int reformat;
     int replacetextplain;
     int decodetext;
     int reformat;
     int replacetextplain;
     int decodetext;

+    char *decodetypes;
+    /* Whether to use CRLF linebreaks, per RFC 2046 Sec. 4.1.1, par.1. */
+    int lf_line_endings;

     char *textcharset;
 } fix_transformations;
 
 int mhfixmsgsbr (CT *, const fix_transformations *, char *);
 static int fix_boundary (CT *, int *);
     char *textcharset;
 } fix_transformations;
 
 int mhfixmsgsbr (CT *, const fix_transformations *, char *);
 static int fix_boundary (CT *, int *);

+static int copy_input_to_output (const char *, const char *);

 static int get_multipart_boundary (CT, char **);
 static int replace_boundary (CT, char *, char *);
 static int fix_types (CT, svector_t, int *);
 static int get_multipart_boundary (CT, char **);
 static int replace_boundary (CT, char *, char *);
 static int fix_types (CT, svector_t, int *);


@@ -106,13 +113,15 @@ static CT build_multipart_alt (CT, CT, int, int);
 static int boundary_in_content (FILE **, char *, const char *);
 static void transfer_noncontent_headers (CT, CT);
 static int set_ct_type (CT, int type, int subtype, int encoding);
 static int boundary_in_content (FILE **, char *, const char *);
 static void transfer_noncontent_headers (CT, CT);
 static int set_ct_type (CT, int type, int subtype, int encoding);

-static int decode_text_parts (CT, int, int *);
+static int decode_text_parts (CT, int, const char *, int *);
+static int should_decode(const char *, const char *, const char *);

 static int content_encoding (CT, const char **);
 static int strip_crs (CT, int *);
 static int convert_charsets (CT, char *, int *);
 static int fix_always (CT, int *);
 static int content_encoding (CT, const char **);
 static int strip_crs (CT, int *);
 static int convert_charsets (CT, char *, int *);
 static int fix_always (CT, int *);

-static int write_content (CT, char *, char *, int, int);
-static int remove_file (char *);
+static int write_content (CT, const char *, char *, int, int);
+static void set_text_ctparams(CT, char *, int);
+static int remove_file (const char *);

 static void report (char *, char *, char *, char *, ...);
 static void pipeser (int);
 
 static void report (char *, char *, char *, char *, ...);
 static void pipeser (int);
 


@@ -135,6 +144,8 @@ main (int argc, char **argv) {
     fx.fixtypes = NULL;
     fx.replacetextplain = 0;
     fx.decodetext = CE_8BIT;
     fx.fixtypes = NULL;
     fx.replacetextplain = 0;
     fx.decodetext = CE_8BIT;

+    fx.decodetypes = "text,application/ics";  /* Default, per man page. */
+    fx.lf_line_endings = 0;

     fx.textcharset = NULL;
 
     if (nmh_init(argv[0], 1)) { return 1; }
     fx.textcharset = NULL;
 
     if (nmh_init(argv[0], 1)) { return 1; }


@@ -179,6 +190,17 @@ main (int argc, char **argv) {
             case NDECODETEXTSW:
                 fx.decodetext = 0;
                 continue;
             case NDECODETEXTSW:
                 fx.decodetext = 0;
                 continue;

+            case DECODETYPESW:
+                if (! (cp = *argp++)  ||  *cp == '-')
+                    adios (NULL, "missing argument to %s", argp[-2]);
+                fx.decodetypes = cp;
+                continue;
+            case CRLFLINEBREAKSSW:
+                fx.lf_line_endings = 0;
+                continue;
+            case NCRLFLINEBREAKSSW:
+                fx.lf_line_endings = 1;
+                continue;

             case TEXTCHARSETSW:
                 if (! (cp = *argp++) || (*cp == '-' && cp[1]))
                     adios (NULL, "missing argument to %s", argp[-2]);
             case TEXTCHARSETSW:
                 if (! (cp = *argp++) || (*cp == '-' && cp[1]))
                     adios (NULL, "missing argument to %s", argp[-2]);


@@ -325,7 +347,23 @@ main (int argc, char **argv) {
         }
         ctp = cts;
 
         }
         ctp = cts;
 

-        if ((ct = parse_mime (file))) { *ctp++ = ct; }
+        if ((ct = parse_mime (file))) {
+            set_text_ctparams(ct, fx.decodetypes, fx.lf_line_endings);
+            *ctp++ = ct;
+        } else {
+            advise (NULL, "unable to parse message from file %s", file);
+            status = NOTOK;
+
+            /* If there's an outfile, pass the input message unchanged, so the message won't
+               get dropped from a pipeline. */
+            if (outfile) {
+                /* Something went wrong.  Output might be expected, such as if this were run
+                   as a filter.  Just copy the input to the output. */
+                if (copy_input_to_output (file, outfile) != OK) {
+                    advise (NULL, "unable to copy message to %s, it might be lost\n", outfile);
+                }
+            }
+        }

     } else {
         /*
          * message(s) are coming from a folder
     } else {
         /*
          * message(s) are coming from a folder


@@ -366,7 +404,25 @@ main (int argc, char **argv) {
                 char *msgnam;
 
                 msgnam = m_name (msgnum);
                 char *msgnam;
 
                 msgnam = m_name (msgnum);

-                if ((ct = parse_mime (msgnam))) { *ctp++ = ct; }
+                if ((ct = parse_mime (msgnam))) {
+                    set_text_ctparams(ct, fx.decodetypes, fx.lf_line_endings);
+                    *ctp++ = ct;
+                } else {
+                    advise (NULL, "unable to parse message %s", msgnam);
+                    status = NOTOK;
+
+                    /* If there's an outfile, pass the input message unchanged, so the message won't
+                       get dropped from a pipeline. */
+                    if (outfile) {
+                        /* Something went wrong.  Output might be expected, such as if this were run
+                           as a filter.  Just copy the input to the output. */
+                        const char *input_filename = path (msgnam, TFILE);
+
+                        if (copy_input_to_output (input_filename, outfile) != OK) {
+                            advise (NULL, "unable to copy message to %s, it might be lost\n", outfile);
+                        }
+                    }
+                }

             }
         }
 
             }
         }
 


@@ -445,7 +501,7 @@ mhfixmsgsbr (CT *ctp, const fix_transformations *fx, char *outfile) {
             ensure_text_plain (ctp, NULL, &message_mods, fx->replacetextplain);
     }
     if (status == OK  &&  fx->decodetext) {
             ensure_text_plain (ctp, NULL, &message_mods, fx->replacetextplain);
     }
     if (status == OK  &&  fx->decodetext) {

-        status = decode_text_parts (*ctp, fx->decodetext, &message_mods);
+        status = decode_text_parts (*ctp, fx->decodetext, fx->decodetypes, &message_mods);

     }
     if (status == OK  &&  fx->textcharset != NULL) {
         status = convert_charsets (*ctp, fx->textcharset, &message_mods);
     }
     if (status == OK  &&  fx->textcharset != NULL) {
         status = convert_charsets (*ctp, fx->textcharset, &message_mods);


@@ -473,19 +529,9 @@ mhfixmsgsbr (CT *ctp, const fix_transformations *fx, char *outfile) {
         /* Something went wrong.  Output might be expected, such
            as if this were run as a filter.  Just copy the input
            to the output. */
         /* Something went wrong.  Output might be expected, such
            as if this were run as a filter.  Just copy the input
            to the output. */

-        int in = open (input_filename, O_RDONLY);
-        int out = strcmp (outfile, "-")
-            ?  open (outfile, O_WRONLY | O_CREAT, m_gmprot ())
-            :  STDOUT_FILENO;
-
-        if (in != -1  &&  out != -1) {
-            cpydata (in, out, input_filename, outfile);
-        } else {
-            status = NOTOK;
+        if (copy_input_to_output (input_filename, outfile) != OK) {
+            advise (NULL, "unable to copy message to %s, it might be lost\n", outfile);

         }
         }

-
-        close (out);
-        close (in);

     }
 
     if (modify_inplace) {
     }
 
     if (modify_inplace) {


@@ -500,6 +546,29 @@ mhfixmsgsbr (CT *ctp, const fix_transformations *fx, char *outfile) {
 }
 
 
 }
 
 

+/* Copy input message to output.  Assumes not modifying in place, so this
+   might be running as part of a pipeline. */
+static int
+copy_input_to_output (const char *input_filename, const char *output_filename) {
+    int in = open (input_filename, O_RDONLY);
+    int out = strcmp (output_filename, "-")
+        ?  open (output_filename, O_WRONLY | O_CREAT, m_gmprot ())
+        :  STDOUT_FILENO;
+    int status = OK;
+
+    if (in != -1  &&  out != -1) {
+        cpydata (in, out, input_filename, output_filename);
+    } else {
+        status = NOTOK;
+    }
+
+    close (out);
+    close (in);
+
+    return status;
+}
+
+

 static int
 fix_boundary (CT *ct, int *message_mods) {
     struct multipart *mp;
 static int
 fix_boundary (CT *ct, int *message_mods) {
     struct multipart *mp;


@@ -556,7 +625,16 @@ fix_boundary (CT *ct, int *message_mods) {
                 }
 
                 free (part_boundary);
                 }
 
                 free (part_boundary);

+            } else {
+                /* Couldn't fix the boundary.  Report failure so that mhfixmsg
+                   doesn't modify the message. */
+                status = NOTOK;

             }
             }

+        } else {
+            /* No multipart struct, even though the content type is
+               CT_MULTIPART.  Report failure so that mhfixmsg doesn't modify
+               the message. */
+            status = NOTOK;

         }
     }
 
         }
     }
 


@@ -1318,8 +1396,7 @@ insert_into_new_mp_alt (CT *ct, int *message_mods) {
         CT mp_alt = build_multipart_alt (*ct, tp_part, CT_MULTIPART,
                                          MULTI_ALTERNATE);
         if (mp_alt) {
         CT mp_alt = build_multipart_alt (*ct, tp_part, CT_MULTIPART,
                                          MULTI_ALTERNATE);
         if (mp_alt) {

-            struct multipart *mp =
-                (struct multipart *) mp_alt->c_ctparams;
+            struct multipart *mp = (struct multipart *) mp_alt->c_ctparams;

 
             if (mp  &&  mp->mp_parts) {
                 mp->mp_parts->mp_part = tp_part;
 
             if (mp  &&  mp->mp_parts) {
                 mp->mp_parts->mp_part = tp_part;


@@ -1340,7 +1417,7 @@ insert_into_new_mp_alt (CT *ct, int *message_mods) {
             status = NOTOK;
         }
     } else {
             status = NOTOK;
         }
     } else {

-        status = NOTOK;
+        /* Not an error if text/plain couldn't be built. */

     }
 
     return status;
     }
 
     return status;


@@ -1748,11 +1825,39 @@ set_ct_type (CT ct, int type, int subtype, int encoding) {
 
 
 static int
 
 
 static int

-decode_text_parts (CT ct, int encoding, int *message_mods) {
+decode_text_parts (CT ct, int encoding, const char *decodetypes, int *message_mods) {

     int status = OK;
     int status = OK;

+    int lf_line_endings = 0;

 
     switch (ct->c_type) {
 
     switch (ct->c_type) {

-    case CT_TEXT:
+    case CT_MULTIPART: {
+        struct multipart *m = (struct multipart *) ct->c_ctparams;
+        struct part *part;
+
+        /* Should check to see if the body for this part is encoded?
+           For now, it gets passed along as-is by InitMultiPart(). */
+        for (part = m->mp_parts; status == OK  &&  part; part = part->mp_next) {
+            status = decode_text_parts (part->mp_part, encoding, decodetypes, message_mods);
+        }
+        break;
+    }
+
+    case CT_MESSAGE:
+        if (ct->c_subtype == MESSAGE_EXTERNAL) {
+            struct exbody *e = (struct exbody *) ct->c_ctparams;
+
+            status = decode_text_parts (e->eb_content, encoding, decodetypes, message_mods);
+        }
+        break;
+
+    default:
+        if (! should_decode(decodetypes, ct->c_ctinfo.ci_type, ct->c_ctinfo.ci_subtype)) {
+            break;
+        }
+
+        lf_line_endings =
+            ct->c_ctparams  &&  ((struct text *) ct->c_ctparams)->lf_line_endings;
+

         switch (ct->c_encoding) {
         case CE_BASE64:
         case CE_QUOTED: {
         switch (ct->c_encoding) {
         case CE_BASE64:
         case CE_QUOTED: {


@@ -1804,7 +1909,9 @@ decode_text_parts (CT ct, int encoding, int *message_mods) {
                             report (NULL, ct->c_partno, ct->c_file, "decode%s",
                                     ct->c_ctline ? ct->c_ctline : "");
                         }
                             report (NULL, ct->c_partno, ct->c_file, "decode%s",
                                     ct->c_ctline ? ct->c_ctline : "");
                         }

-                        strip_crs (ct, message_mods);
+                        if (lf_line_endings) {
+                            strip_crs (ct, message_mods);
+                        }

                     } else {
                         status = NOTOK;
                     }
                     } else {
                         status = NOTOK;
                     }


@@ -1816,39 +1923,49 @@ decode_text_parts (CT ct, int encoding, int *message_mods) {
         }
         case CE_8BIT:
         case CE_7BIT:
         }
         case CE_8BIT:
         case CE_7BIT:

-            strip_crs (ct, message_mods);
+            if (lf_line_endings) {
+                strip_crs (ct, message_mods);
+            }

             break;
         default:
             break;
         }
 
         break;
             break;
         default:
             break;
         }
 
         break;

-
-    case CT_MULTIPART: {
-        struct multipart *m = (struct multipart *) ct->c_ctparams;
-        struct part *part;
-
-        /* Should check to see if the body for this part is encoded?
-           For now, it gets passed along as-is by InitMultiPart(). */
-        for (part = m->mp_parts; status == OK  &&  part; part = part->mp_next) {
-            status = decode_text_parts (part->mp_part, encoding, message_mods);
-        }
-        break;

     }
 
     }
 

-    case CT_MESSAGE:
-        if (ct->c_subtype == MESSAGE_EXTERNAL) {
-            struct exbody *e = (struct exbody *) ct->c_ctparams;
+    return status;
+}

 
 

-            status = decode_text_parts (e->eb_content, encoding, message_mods);
-        }
-        break;

 
 

-    default:
-        break;
+/* Determine if the part with type[/subtype] should be decoded, according to
+   decodetypes (which came from the -decodetypes switch). */
+static int
+should_decode(const char *decodetypes, const char *type, const char *subtype) {
+    /* Quick search for matching type[/subtype] in decodetypes:  bracket
+       decodetypes with commas, then search for ,type, and ,type/subtype, in
+       it. */
+
+    int found_match = 0;
+    char *delimited_decodetypes = concat(",", decodetypes, ",", NULL);
+    char *delimited_type = concat(",", type, ",", NULL);
+
+    if (nmh_strcasestr(delimited_decodetypes, delimited_type)) {
+        found_match = 1;
+    } else if (subtype != NULL) {
+        char *delimited_type_subtype =
+            concat(",", type, "/", subtype, ",", NULL);
+
+        if (nmh_strcasestr(delimited_decodetypes, delimited_type_subtype)) {
+            found_match = 1;
+        }
+        free(delimited_type_subtype);

     }
 
     }
 

-    return status;
+    free(delimited_type);
+    free(delimited_decodetypes);
+
+    return found_match;

 }
 
 
 }
 
 


@@ -2200,7 +2317,7 @@ fix_always (CT ct, int *message_mods) {
 
 
 static int
 
 
 static int

-write_content (CT ct, char *input_filename, char *outfile, int modify_inplace,
+write_content (CT ct, const char *input_filename, char *outfile, int modify_inplace,

                int message_mods) {
     int status = OK;
 
                int message_mods) {
     int status = OK;
 


@@ -2273,12 +2390,50 @@ write_content (CT ct, char *input_filename, char *outfile, int modify_inplace,
 }
 
 
 }
 
 

+/*
+ * parse_mime() does not set lf_line_endings in struct text, so use this function to do it.
+ * It touches the parts the decodetypes identifies.
+ */
+static void
+set_text_ctparams(CT ct, char *decodetypes, int lf_line_endings) {
+    switch (ct->c_type) {
+    case CT_MULTIPART: {
+        struct multipart *m = (struct multipart *) ct->c_ctparams;
+        struct part *part;
+
+        for (part = m->mp_parts; part; part = part->mp_next) {
+            set_text_ctparams(part->mp_part, decodetypes, lf_line_endings);
+        }
+        break;
+    }
+
+    case CT_MESSAGE:
+        if (ct->c_subtype == MESSAGE_EXTERNAL) {
+            struct exbody *e = (struct exbody *) ct->c_ctparams;
+
+            set_text_ctparams(e->eb_content, decodetypes, lf_line_endings);
+        }
+        break;
+
+    default:
+        if (should_decode(decodetypes, ct->c_ctinfo.ci_type, ct->c_ctinfo.ci_subtype)) {
+            if (ct->c_ctparams == NULL) {
+                if ((ct->c_ctparams = (struct text *) mh_xcalloc (1, sizeof (struct text))) == NULL) {
+                    adios (NULL, "out of memory");
+                }
+            }
+            ((struct text *) ct->c_ctparams)->lf_line_endings = lf_line_endings;
+        }
+    }
+}
+
+

 /*
  * If "rmmproc" is defined, call that to remove the file.  Otherwise,
  * use the standard MH backup file.
  */
 static int
 /*
  * If "rmmproc" is defined, call that to remove the file.  Otherwise,
  * use the standard MH backup file.
  */
 static int

-remove_file (char *file) {
+remove_file (const char *file) {

     if (rmmproc) {
         char *rmm_command = concat (rmmproc, " ", file, NULL);
         int status = system (rmm_command);
     if (rmmproc) {
         char *rmm_command = concat (rmmproc, " ", file, NULL);
         int status = system (rmm_command);